Feels like every idea now has “AI” stapled onto it.
Curious what people would focus on if that wasn’t an option, what problems would actually be worth solving?
I still want to build a bitgrid, and see if I can democratize petaflops.
Other things I'd like to build/have include an actual Memex that can ingest HTML, and actually use it to mark up hypertext, logging everything for at least a month before discarding the stuff I don't reference again from the cache.
I did that once! It was a rare opportunity caused by a system whose database was (predictably) corrupted by six months of programming errors. Took 3 days to examine the problem and code the fix and 36 minutes of tense run-time to fix the entire database (followed by 3 days of checking done by the DBA independently). What a rush - can't say I miss it though!
Other things I'd like to build/have include an actual Memex that can ingest HTML, and actually use it to mark up hypertext, logging everything for at least a month before discarding the stuff I don't reference again from the cache.